AsranielJoin Date: 2002-06-03Member: 724Members, Playtest Lead, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow, Subnautica Playtester, Retired Community Developer
edited October 2015
The renderer was not changes, especially not related to anything with g-sync.
So more information would be needed. Can you do the following:
Enter sandbox and check you get the low fps.
open the console and type p_logall
after about 5-10 seconds type p_endlog in the console
Now you will have a file created in your %appdata% Natural Selection 2 folder. You will have to upload this file later.
Now to have a better understanding of whats going on, do your workaround to get 200 fps. Then repeat the previous procedure, that will create a second plog file.
Upload both files somewhere and link them here
Also, make sure you look at the exact same spot, ideally a wall or something in both tests.
The renderer was not changes, especially not related to anything with g-sync.
...Upload both files somewhere and link them here
Also, make sure you look at the exact same spot, ideally a wall or something in both tests.
Link to Zip file with both Win10 and Win7. With Gsync enabled, you'll note Win10 is around 113fps and Win7 is normal at 143-144. All game settings and nVidia settings are the same. Both in the Docking sandbox ready room - same position. Win10 Gsync on Build 275 worked correctly.
I forgot I had updated to nVidia driver version 358.50 (Win10 64bit) for the Battlefront weekend. I now believe this to be the problem...nVidia broke Gsync with 358.50, at least when playing NS2. I reverted back to version 355.98 and I'm back to normal Gsync framerates of 143-144fps at all times when playing NS2. I'll stick with this driver version for now.
I forgot I had updated to nVidia driver version 358.50 (Win10 64bit) for the Battlefront weekend. I now believe this to be the problem...nVidia broke Gsync with 358.50, at least when playing NS2. I reverted back to version 355.98 and I'm back to normal Gsync framerates of 143-144fps at all times when playing NS2. I'll stick with this driver version for now.
We cant help with driver issues I'm afraid. Nothing was changed on the renderer side.
We cant help with driver issues I'm afraid. Nothing was changed on the renderer side.
Of course...this is purely an nVidia issue. The patch (on previous nVidia drivers) seems okay with my limited testing. I still get the sound squeak right when I start the game as has been reported. That's probably a Win10 and sound card/driver issue I imagine. Thanks for the work on the patch CDT!
@Webtrance Bit off-topic, but do you normally play NS2 with G-Sync turned on? Any advantages using G-Sync over ULMB? I haven't touched G-Sync for FPS games since, although it alleviates large FPS fluctuations by smoothing them out, it results in horrendous motion blur. The frame rate has been optimized in NS2 so much since beta that I find no need for G-Sync. I strictly use ULMB for the motion clarity since frame rates are high and fairly stable until end game.
Yeah looking at your plog i'm almost certain its the fault of nvidia.. i'll look a little further though.
You say it worked with 275. Are you sure you had the exact same driver?
I'm sure I had the Battlefront drivers (358.50) installed on 275. However, I went about 5-6 days without playing NS2 after installing 358.50 and then 276 came out. Naturally, I thought that 276 broke Gsync, when it was 358.50 that actually broke it. Like I said though, reverting back to version 355.98 made Gsync work again.
@Webtrance Bit off-topic, but do you normally play NS2 with G-Sync turned on? Any advantages using G-Sync over ULMB? I haven't touched G-Sync for FPS games since, although it alleviates large FPS fluctuations by smoothing them out, it results in horrendous motion blur. The frame rate has been optimized in NS2 so much since beta that I find no need for G-Sync. I strictly use ULMB for the motion clarity since frame rates are high and fairly stable until end game.
I play with Gsync on all the time. ULMB is too dark for me, lacks color, microstutters, and tears. The only way I'd play with ULMB is to vcync it to reduce microstutter and tears. The problem with vsync at 120hz is if you drop below 120fps vsync will turn off, assuming you're using adaptive vsync. The "jolt" in loosing a smooth 120fps vsync'd by droping just below 120fps (as adaptive vsync turns off) is too much for me to bear. Besides, UMLB will only work up to 120hz, so it sort of gimps your nice 144hz monitor in that respect.
Gsync is where it at imo. I typically sustain 143-144fps until later game, where I'll drop to the 120's-130s. Keep in mind that I'm still Gsync'd between 120fps and 144fps. Gsync enabled will automatically turn on vsync in the nVidia control panel. Of course, make sure "Turn on Veritcal Sync" is OFF in NS2 options.
The bottom line, I use Gsync because it's like vsync for me, less the associated input lag. It completely "smooths" out the game for me. Before Gsync, I would framerate limit to 140fps (4fps below my 144hz monitor). Reason: Capping 4fps below my 144hz refresh rate worked better than capping at 2fps below or 142fps. A steady 140fps kept microstutter and tearing to a minimum. I see many folks stream this game and they're at 180-200fps at the start of the game. I'll bet if they actually take care to notice when panning back and forth with smooth mouse movements, they'll see tearing and will notice a "judder" of micro stuttering. Most folks just don't know any better or don't care. Now, let's get them to vsync their game on a 144hz monitor. It's perfectly smooth and nice. But, drop below 144fps and adaptive vsync turns off and your left with 100-143fps of non-vsync gameplay with tearing and microstutter. Enter Gsync....now you can play "vsync'd", if you will, all the way down to 30fps, all the while minimizing mouse input lag. Of course, even with Gsync on and getting 60fps, for example, you're still going to notice a clear difference in your mouse response. So for me, it's Gsync all the way!
ZavaroTucson, ArizonaJoin Date: 2005-02-14Member: 41174Members, Super Administrators, NS2 Playtester, Squad Five Blue, Squad Five Silver, NS2 Map Tester, Reinforced - Shadow, WC 2013 - Silver, Subnautica Playtester, NS2 Community Developer, Pistachionauts
I wish I had a Gsync monitor. I'm not sure how many devs or PTs have them at this point, but I think its zero. A shame those game ready drivers caused lag. I was having a similar issue and have since reverted, giving me an addition 20-ish frames.
I wish I had a Gsync monitor. I'm not sure how many devs or PTs have them at this point, but I think its zero. A shame those game ready drivers caused lag. I was having a similar issue and have since reverted, giving me an addition 20-ish frames.
Gsync is so nice...it's makes NS2 and other games actually enjoyable for me. Before Gsync, I tried everything to minimize the slight microstutter and "bumpy/juddery" movements I was seeing when I would mouse side to side. Playing at 60hz is like a slideshow to me and I just can't do it. I started NS2 on my old 60hz monitor and was okay. Then I overclocked my monitor to 75hz and it was like, wow! Then I got a Benq 144hz monitor and I was like SWEET! But, I still got slight tearing and microstutter. So I started framerate capping below the 144hz refresh rate. This smoothed it all out...enough to be happy for about 1.5 years. Then Gsync...it was like when I tried playing games vsynced. However, I never played with vsync because of the inherent problems with it...mouse lag in particular. Also, the on/off nature of adaptive vsync. It was like vsync was on then off, on then off, etc. This was very annoying and ruined the game experience. Gsync solves all my problems in terms of having smooth game play at all times. Since I normally stay at 120fps and above, I don't notice any mouse input lag either.
Note: I tried the Asus ROG Swift PG278Q and the Acer XB270HU, both 1440p. Both are excellent. BUT!!! The goal is to sustain as close to 144fps throughout your entire game/match to benefit from ultra-smooth Gsync game play. Playing at 1440p with high textures and other graphic quality settings can be taxing on your system. Hence, I'd recommend gaming at 1080p. I use the Acer XB270H, 1920x1080 27". I don't feel that QHD or 4K gaming is the way to go these days with limited CPU/GPU power. Unless, of course, you're happy with 40, 60-100fps at times. For me it's all about framerate and maximizing my FPS. In NS2, I still play with high textures, high light quality, and 16x anisotropic filtering. That's it. It keeps me at higher framerates to enjoy Gsync and stay as close as I can to 144fps.
DC_DarklingJoin Date: 2003-07-10Member: 18068Members, Constellation, Squad Five Blue, Squad Five Silver
I also have a gsync monitor and damn its awesome.
definitely play your games in gsync first.
No stutters, no spikes, no tearing. Its the way to go. Gsync also plays nice with low fps.
ULMB instead of gsync is nice if the above is not a issue. Usually you need very high fps to sustain ULMB.
Do note that if you drop over your monitors max refresh rate, gsync disables. It then either goes vsync or not depending on settings.
So making sure in games that you do not breach 144fps would help with gsync. Max out those settings. :P
Interesting. I have become a big fan of ULMB. 144hz is almost not a noticeable different from 60hz for me, but ULMB was night and day. So it seems the tradeoff between ULMB and Gsync is clarity vs smoothness. Do you gsync users agree with that it is a tradeoff between clarity and smoothness?
ULMB makes is so you have no after images.
gsync makes it so you have less tearing and stutter.
Its a personal thing what annoys you more. Depends on your eyes also.
So I guess we can agree.
I have yet to experience a gsync monitor, and I already did not think I wanted one. I hardly notice micro stutters unless they are really bad. I appreciate you guys describing your experiences with both.
DC_DarklingJoin Date: 2003-07-10Member: 18068Members, Constellation, Squad Five Blue, Squad Five Silver
No problem, each their own.
I have a monitor capable of both and I love to use gsync in many games and ULMB on other stuff like desktop. ^^
For me ULMB is easier on the eyes but in fast games the stutters were even harder to take.
@Webtrance@DC_Darkling Thanks for your guys' input. Really surprised to hear that you don't mind Gsync's motion blur. I just prefer ULMB since motion blur is more distracting to me than micro stutters, which I do notice in NS2. I agree that ULMB's dull, washed out colors are a big downside. They can be somewhat improved by increasing gamma and digital vibrance in Nvidia control panel, which I've gotten used to doing. I'll have to test Gsync again tonight.
DC_DarklingJoin Date: 2003-07-10Member: 18068Members, Constellation, Squad Five Blue, Squad Five Silver
ULMB is a lot more doable when calibrated. Not superb, but much better. Like any monitor when calibrated.
For me its more stutters and tearing, rather then blur.
Do not get me wrong, I definitely see blurring also in many things, just the other stuff more.
IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
edited October 2015
@Karnaj Gsync's motion blur is still 60% less than the fastest responding 60 hz LCD (16.8 ms)
True, it's 6 ms instead of 1.4 -2.8 ms (dependent on hz) from using lightboost, which is close to matching a CRT monitor in persistence, but it's still a hell of an improvement and imo worth it when you weigh the benefits between them.
Lightboost is only improving fluidity, whereas Gsync improves the fluidity (albeit not to the same degree) and much more, from stuttering to tearing to a wider applicable range of FPS and it does not have the flicker issues, brightness issues, degradation of color issues that Lightboost suffers from.
Even though better persistence assists in NS2, Gsync is preferable in this game solely because of the HIGH fluctuation in FPS that occurs that is seemingly impossible to address thus far.
DC_DarklingJoin Date: 2003-07-10Member: 18068Members, Constellation, Squad Five Blue, Squad Five Silver
What ironhorse said with a slight addition.
Lightboost is basicly version 1. ULMB is version 2 and has pretty much noone of the ugly drawbacks. I see no flickers for example. (but that may vary from person to person.)
So I played using Gsync again for an hour or two last night and, although the smooth motion, lack of stutters, and the color improvement are nice, as expected, I really can't stand the motion blur. Fading/skulking were the biggest problems for me and it became much harder tracking juking marines. I went back to ULMB and the lack of motion blur outweighs the micro stutters for me.
Maybe I'm just too used to playing this way. I had a different monitor with LightBoost and then I used this ASUS VG248QE for the longest time until the Gsync upgrade kit came out. My only gripe with ULMB on my monitor is that the strobe length can't be adjusted and it's stuck at 2 ms pixel persistence (LightBoost @ 50%). Not a huge difference from 1.4~ ms persistence (LightBoost @ 10%), but still.
FrozenNew York, NYJoin Date: 2010-07-02Member: 72228Members, Constellation
I turn blur off in everything, as well as depth of field, it just sort of makes me sick. That's not happening to me with Gsync @120+fps. Are you sure it's from Gsync?
I turn blur off in everything, as well as depth of field, it just sort of makes me sick. That's not happening to me with Gsync @120+fps. Are you sure it's from Gsync?
I mean display motion blur, not in-game settings. Gsync and ULMB/LightBoost are like polar opposites in their advantages and drawbacks. ULMB/LightBoost is designed to significantly reduce motion blur, but it doesn't smooth out variable frame rates or fix stutters and screen tearing like Gsync does. On the other hand, Gsync doesn't do anything to reduce motion blur, so I've been seeing a lot of it while testing it compared to ULMB.
I can't seem to find any measurements of pixel persistence while using Gsync, but I have a feeling it's around 7 ms, considering this is the measurement for a 144 Hz monitor.
DC_DarklingJoin Date: 2003-07-10Member: 18068Members, Constellation, Squad Five Blue, Squad Five Silver
Exactly as Karnaj describes.
The difference is there. Some people are just more susceptible then others.
I feel the problem of both the eyes own blurring as well as the stutters, but for me the g-sync in general works better.
But for most of my monitors use, its in anti blur.
Its why I still have not upgraded my nvidia drivers. In 'new/latest' drivers you can not have ULMB on desktop while retaining g-sycn usage in games, and nvidia told me its not coming back in any short period.
/sads.
@IronHorse I believe your statement is misleading, though. A 144 Hz refresh rate has 6.9 ms~ pixel persistence with or without Gsync functionality. In this case, it's the higher refresh rate of 144 Hz that causes less motion blur (6.9 ms) compared to 60 Hz (16.7 ms). Gsync itself does not reduce motion blur and I haven't been able to find evidence of it doing so over at Blur Busters.
I feel as if we're rehashing and sometimes misunderstanding what Mark Rejhon at Blur Busters has already investigated. Hopefully, it won't be too long until Gsync and ULMB can be combined together.
Comments
So more information would be needed. Can you do the following:
Enter sandbox and check you get the low fps.
open the console and type p_logall
after about 5-10 seconds type p_endlog in the console
Now you will have a file created in your %appdata% Natural Selection 2 folder. You will have to upload this file later.
Now to have a better understanding of whats going on, do your workaround to get 200 fps. Then repeat the previous procedure, that will create a second plog file.
Upload both files somewhere and link them here
Also, make sure you look at the exact same spot, ideally a wall or something in both tests.
Link to Zip file with both Win10 and Win7. With Gsync enabled, you'll note Win10 is around 113fps and Win7 is normal at 143-144. All game settings and nVidia settings are the same. Both in the Docking sandbox ready room - same position. Win10 Gsync on Build 275 worked correctly.
https://www.dropbox.com/s/pu8ei17i3b49nrg/client-1017-083023plogs_Win7-10.zip?dl=0
We cant help with driver issues I'm afraid. Nothing was changed on the renderer side.
Of course...this is purely an nVidia issue. The patch (on previous nVidia drivers) seems okay with my limited testing. I still get the sound squeak right when I start the game as has been reported. That's probably a Win10 and sound card/driver issue I imagine. Thanks for the work on the patch CDT!
You say it worked with 275. Are you sure you had the exact same driver?
I'm sure I had the Battlefront drivers (358.50) installed on 275. However, I went about 5-6 days without playing NS2 after installing 358.50 and then 276 came out. Naturally, I thought that 276 broke Gsync, when it was 358.50 that actually broke it. Like I said though, reverting back to version 355.98 made Gsync work again.
I play with Gsync on all the time. ULMB is too dark for me, lacks color, microstutters, and tears. The only way I'd play with ULMB is to vcync it to reduce microstutter and tears. The problem with vsync at 120hz is if you drop below 120fps vsync will turn off, assuming you're using adaptive vsync. The "jolt" in loosing a smooth 120fps vsync'd by droping just below 120fps (as adaptive vsync turns off) is too much for me to bear. Besides, UMLB will only work up to 120hz, so it sort of gimps your nice 144hz monitor in that respect.
Gsync is where it at imo. I typically sustain 143-144fps until later game, where I'll drop to the 120's-130s. Keep in mind that I'm still Gsync'd between 120fps and 144fps. Gsync enabled will automatically turn on vsync in the nVidia control panel. Of course, make sure "Turn on Veritcal Sync" is OFF in NS2 options.
The bottom line, I use Gsync because it's like vsync for me, less the associated input lag. It completely "smooths" out the game for me. Before Gsync, I would framerate limit to 140fps (4fps below my 144hz monitor). Reason: Capping 4fps below my 144hz refresh rate worked better than capping at 2fps below or 142fps. A steady 140fps kept microstutter and tearing to a minimum. I see many folks stream this game and they're at 180-200fps at the start of the game. I'll bet if they actually take care to notice when panning back and forth with smooth mouse movements, they'll see tearing and will notice a "judder" of micro stuttering. Most folks just don't know any better or don't care. Now, let's get them to vsync their game on a 144hz monitor. It's perfectly smooth and nice. But, drop below 144fps and adaptive vsync turns off and your left with 100-143fps of non-vsync gameplay with tearing and microstutter. Enter Gsync....now you can play "vsync'd", if you will, all the way down to 30fps, all the while minimizing mouse input lag. Of course, even with Gsync on and getting 60fps, for example, you're still going to notice a clear difference in your mouse response. So for me, it's Gsync all the way!
Gsync is so nice...it's makes NS2 and other games actually enjoyable for me. Before Gsync, I tried everything to minimize the slight microstutter and "bumpy/juddery" movements I was seeing when I would mouse side to side. Playing at 60hz is like a slideshow to me and I just can't do it. I started NS2 on my old 60hz monitor and was okay. Then I overclocked my monitor to 75hz and it was like, wow! Then I got a Benq 144hz monitor and I was like SWEET! But, I still got slight tearing and microstutter. So I started framerate capping below the 144hz refresh rate. This smoothed it all out...enough to be happy for about 1.5 years. Then Gsync...it was like when I tried playing games vsynced. However, I never played with vsync because of the inherent problems with it...mouse lag in particular. Also, the on/off nature of adaptive vsync. It was like vsync was on then off, on then off, etc. This was very annoying and ruined the game experience. Gsync solves all my problems in terms of having smooth game play at all times. Since I normally stay at 120fps and above, I don't notice any mouse input lag either.
Note: I tried the Asus ROG Swift PG278Q and the Acer XB270HU, both 1440p. Both are excellent. BUT!!! The goal is to sustain as close to 144fps throughout your entire game/match to benefit from ultra-smooth Gsync game play. Playing at 1440p with high textures and other graphic quality settings can be taxing on your system. Hence, I'd recommend gaming at 1080p. I use the Acer XB270H, 1920x1080 27". I don't feel that QHD or 4K gaming is the way to go these days with limited CPU/GPU power. Unless, of course, you're happy with 40, 60-100fps at times. For me it's all about framerate and maximizing my FPS. In NS2, I still play with high textures, high light quality, and 16x anisotropic filtering. That's it. It keeps me at higher framerates to enjoy Gsync and stay as close as I can to 144fps.
Cheers!
definitely play your games in gsync first.
No stutters, no spikes, no tearing. Its the way to go. Gsync also plays nice with low fps.
ULMB instead of gsync is nice if the above is not a issue. Usually you need very high fps to sustain ULMB.
Do note that if you drop over your monitors max refresh rate, gsync disables. It then either goes vsync or not depending on settings.
So making sure in games that you do not breach 144fps would help with gsync. Max out those settings. :P
You can not mix both techniques, sadly.
So yes, ns2 in gsync if you can. 100% agree.
gsync makes it so you have less tearing and stutter.
Its a personal thing what annoys you more. Depends on your eyes also.
So I guess we can agree.
I have yet to experience a gsync monitor, and I already did not think I wanted one. I hardly notice micro stutters unless they are really bad. I appreciate you guys describing your experiences with both.
I have a monitor capable of both and I love to use gsync in many games and ULMB on other stuff like desktop. ^^
For me ULMB is easier on the eyes but in fast games the stutters were even harder to take.
For me its more stutters and tearing, rather then blur.
Do not get me wrong, I definitely see blurring also in many things, just the other stuff more.
True, it's 6 ms instead of 1.4 -2.8 ms (dependent on hz) from using lightboost, which is close to matching a CRT monitor in persistence, but it's still a hell of an improvement and imo worth it when you weigh the benefits between them.
Lightboost is only improving fluidity, whereas Gsync improves the fluidity (albeit not to the same degree) and much more, from stuttering to tearing to a wider applicable range of FPS and it does not have the flicker issues, brightness issues, degradation of color issues that Lightboost suffers from.
Even though better persistence assists in NS2, Gsync is preferable in this game solely because of the HIGH fluctuation in FPS that occurs that is seemingly impossible to address thus far.
Lightboost is basicly version 1. ULMB is version 2 and has pretty much noone of the ugly drawbacks. I see no flickers for example. (but that may vary from person to person.)
Maybe I'm just too used to playing this way. I had a different monitor with LightBoost and then I used this ASUS VG248QE for the longest time until the Gsync upgrade kit came out. My only gripe with ULMB on my monitor is that the strobe length can't be adjusted and it's stuck at 2 ms pixel persistence (LightBoost @ 50%). Not a huge difference from 1.4~ ms persistence (LightBoost @ 10%), but still.
I can't seem to find any measurements of pixel persistence while using Gsync, but I have a feeling it's around 7 ms, considering this is the measurement for a 144 Hz monitor.
The difference is there. Some people are just more susceptible then others.
I feel the problem of both the eyes own blurring as well as the stutters, but for me the g-sync in general works better.
But for most of my monitors use, its in anti blur.
Its why I still have not upgraded my nvidia drivers. In 'new/latest' drivers you can not have ULMB on desktop while retaining g-sycn usage in games, and nvidia told me its not coming back in any short period.
/sads.
I feel as if we're rehashing and sometimes misunderstanding what Mark Rejhon at Blur Busters has already investigated. Hopefully, it won't be too long until Gsync and ULMB can be combined together.